Database problems(solved)

pmtechwl

Verified User
Joined
Mar 8, 2016
Messages
8
I used Custombuild to update MariaDB from 10.3.29 to 10.3.30 and after that I am unable to login to any of my wordpress websites.
I tried resetting password but i get "Your password reset link appears to be invalid. Please request a new link below."
My server is a VPS CentOS 7,
Apache 2.4.48Running
DirectAdmin 1.62.4Running
Exim 4.92.3Running
MariaDB 10.4.20Running
Named 9.11.4Running
sshd Running
Nginx 1.21.0Running
dovecot 2.3.15 (0503334ab1)Running
pure-ftpd 1.0.49Running
Php 7.4.20Installed
Php 7.1.33Installed (php2)
Php 7.3.28Installed (php3)
Php 5.6.40Installed (php4)

ANyone on this issue? I tried going back to mariaDB 10.2 but no luck, upgraded to 10.4 also no luck.

Help?
 
Hi Brent,

MYSQLD is running, database has connections, but cannot login to admin of any wordpress site.

Wordpress sites frontends are up.

I installed a webshop last week and that is down and shows: Warning: mysqli::query(): (HY000/1): Can't create/write to file '/tmp/#sql_4f2b_0.MAI' (Errcode: 2 "No such file or directory") in /home/blacklabs/domains/blacklabs.eu/public_html/system/library/db/mysqli.php on line 19

PHPmyadmin of one of the sites says "no permissions". These sites have been running fine for more than a year until yesterday.
 
Did you check that you have a /tmp/ directory like the error message indicates that you don't have? If you have somehow managed to delete it, you must recreate it again.
 
I used custombuilt to update, and have not touched anything else or deleted. /tmp/ folders are are there.
 
ls -ld /tmp
drwxrwxrwt. 17 root root 4096 1 jul 12:44 /tmp

df -h /tmp
Bestandssysteem Grootte Gebruikt Besch Geb% Aangekoppeld op
/dev/mapper/vg_directadmin-lv_tmp 1014M 33M 982M 4% /tmp

df -hi /tmp
Bestandssysteem I-nodes IGebr IVrij IGeb% Aangekoppeld op
/dev/mapper/vg_directadmin-lv_tmp 512K 40 512K 1% /tmp
 
It is possible that your 1GB /tmp partition fills up quickly if there are multiple queries running that need temp space on disk. I would monitor it in close to realtime to see if there are any spikes filling it up. It's also possible that MySQL is using a systemd controlled tmp directory, which could possibly cause some issues. See the various answers/comments here for other possible solutions: https://stackoverflow.com/questions...to-file-tmp-sql-3c6-0-myi-errcode-2-what-does
 
there are 4 websites with no high volume on it, so that seems not to be the issue, and it worked before the custombuilt update of MariaDB.....
 
no changes made, just Update and then problems. I checked all files and permissions and no go. I also downgraded MariaDB, no resolution, and finally did a server reboot and now it works again! Very strange but at least i can manage the websites again!
 
no changes made, just Update and then problems. I checked all files and permissions and no go. I also downgraded MariaDB, no resolution, and finally did a server reboot and now it works again! Very strange but at least i can manage the websites again!
DOWNGRADE MariaDB is very dangerous! Bad experience with, you should have good backups before upgrading MARIASB MYSQL, and downgrade even more.

THE files i mentioned could it be that there was a kind of renamed my.cnf (new) with newer settinsg or those for other config files?
 
If a reboot resolved it, did you try a restart of MySQL first? If you encountered the systemd tmp directory issue, that would've resolved it. It is likely it will happen again though, so I'd look into it if I were you.
 
I have VPS Backups, but was eager to see how to resolve. finally a reboot of the VPS fixed it.
Everything came after custombuild updates, so be extra carefull with that.
 
I tried averything top down, restart MYSQL, restart Ngnx and Apache, rebooting the VPS before restoring frmom VPS Backup solved the issue!

my.cnf file had not replaced or altered. Thats the wierd thing.
 
Back
Top